{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:IXS4NSLLGB7SMYN5KOKPR5DACB","short_pith_number":"pith:IXS4NSLL","schema_version":"1.0","canonical_sha256":"45e5c6c96b307f2661bd5394f8f460106cc1856caf299afd390bb8b29243651a","source":{"kind":"arxiv","id":"2501.05966","version":2},"attestation_state":"computed","paper":{"title":"Towards Early Prediction of Self-Supervised Speech Model Performance","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.LG","eess.AS"],"primary_cat":"cs.SD","authors_text":"Lucas Maison, Marco Dinarelli, Ryan Whetten, Titouan Parcollet, Yannick Est\\`eve","submitted_at":"2025-01-10T13:49:09Z","abstract_excerpt":"In Self-Supervised Learning (SSL), pre-training and evaluation are resource intensive. In the speech domain, current indicators of the quality of SSL models during pre-training, such as the loss, do not correlate well with downstream performance. Consequently, it is often difficult to gauge the final downstream performance in a cost efficient manner during pre-training. In this work, we propose unsupervised efficient methods that give insights into the quality of the pre-training of SSL speech models, namely, measuring the cluster quality and rank of the embeddings of the SSL model. Results sh"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2501.05966","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SD","submitted_at":"2025-01-10T13:49:09Z","cross_cats_sorted":["cs.CL","cs.LG","eess.AS"],"title_canon_sha256":"4ac3553f3d757c6e060ccd24370fe2660ffeae64073af528f4fdebbdd948a037","abstract_canon_sha256":"bece7d6433ba4808b346096cfff2f481b4e9021d9b340ce62a9c81002d4461a3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:13:18.542184Z","signature_b64":"XfdH0g53AWE5LraP1Hk/+j6o1gwMhhAa2Suh/VBB1eYwimNDGhj9TSQanIkZThcWXTUPWkXal3TN8abH6T82Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"45e5c6c96b307f2661bd5394f8f460106cc1856caf299afd390bb8b29243651a","last_reissued_at":"2026-07-05T11:13:18.541710Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:13:18.541710Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Towards Early Prediction of Self-Supervised Speech Model Performance","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.LG","eess.AS"],"primary_cat":"cs.SD","authors_text":"Lucas Maison, Marco Dinarelli, Ryan Whetten, Titouan Parcollet, Yannick Est\\`eve","submitted_at":"2025-01-10T13:49:09Z","abstract_excerpt":"In Self-Supervised Learning (SSL), pre-training and evaluation are resource intensive. In the speech domain, current indicators of the quality of SSL models during pre-training, such as the loss, do not correlate well with downstream performance. Consequently, it is often difficult to gauge the final downstream performance in a cost efficient manner during pre-training. In this work, we propose unsupervised efficient methods that give insights into the quality of the pre-training of SSL speech models, namely, measuring the cluster quality and rank of the embeddings of the SSL model. Results sh"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.05966","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2501.05966/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2501.05966","created_at":"2026-07-05T11:13:18.541765+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.05966v2","created_at":"2026-07-05T11:13:18.541765+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.05966","created_at":"2026-07-05T11:13:18.541765+00:00"},{"alias_kind":"pith_short_12","alias_value":"IXS4NSLLGB7S","created_at":"2026-07-05T11:13:18.541765+00:00"},{"alias_kind":"pith_short_16","alias_value":"IXS4NSLLGB7SMYN5","created_at":"2026-07-05T11:13:18.541765+00:00"},{"alias_kind":"pith_short_8","alias_value":"IXS4NSLL","created_at":"2026-07-05T11:13:18.541765+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2501.05966","citing_title":"Towards Early Prediction of Self-Supervised Speech Model Performance","ref_index":2,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IXS4NSLLGB7SMYN5KOKPR5DACB","json":"https://pith.science/pith/IXS4NSLLGB7SMYN5KOKPR5DACB.json","graph_json":"https://pith.science/api/pith-number/IXS4NSLLGB7SMYN5KOKPR5DACB/graph.json","events_json":"https://pith.science/api/pith-number/IXS4NSLLGB7SMYN5KOKPR5DACB/events.json","paper":"https://pith.science/paper/IXS4NSLL"},"agent_actions":{"view_html":"https://pith.science/pith/IXS4NSLLGB7SMYN5KOKPR5DACB","download_json":"https://pith.science/pith/IXS4NSLLGB7SMYN5KOKPR5DACB.json","view_paper":"https://pith.science/paper/IXS4NSLL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.05966&json=true","fetch_graph":"https://pith.science/api/pith-number/IXS4NSLLGB7SMYN5KOKPR5DACB/graph.json","fetch_events":"https://pith.science/api/pith-number/IXS4NSLLGB7SMYN5KOKPR5DACB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IXS4NSLLGB7SMYN5KOKPR5DACB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IXS4NSLLGB7SMYN5KOKPR5DACB/action/storage_attestation","attest_author":"https://pith.science/pith/IXS4NSLLGB7SMYN5KOKPR5DACB/action/author_attestation","sign_citation":"https://pith.science/pith/IXS4NSLLGB7SMYN5KOKPR5DACB/action/citation_signature","submit_replication":"https://pith.science/pith/IXS4NSLLGB7SMYN5KOKPR5DACB/action/replication_record"}},"created_at":"2026-07-05T11:13:18.541765+00:00","updated_at":"2026-07-05T11:13:18.541765+00:00"}